I first replicate Donohue and Levitt's results for violent and property crime arrest rates. I apply their data and specification to an analysis of age-specific homicide rates and murder arrest rates. The coefficients on the abortion rate have the wrong sign for two of the four measures of crime...